Extraction, identification, and quantification of antioxidant phenolics from hazelnut (Corylus avellana L.) shells

Food Chemistry 244 (2018) 7–15

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Extraction, identification, and quantification of antioxidant phenolics from hazelnut (Corylus avellana L.) shells Bo Yuana,b, Mei Lua, Kent M. Eskridgec, Loren D. Isomb, Milford A. Hannaa,b, a b c

MARK

⁎

Department of Food Science and Technology, University of Nebraska-Lincoln, 1901 North 21st Street, Lincoln, NE 68588-6205, USA Industrial Agricultural Products Center, University of Nebraska-Lincoln, 208 L.W. Chase Hall, Lincoln, NE 68583-0726, USA Department of Statistics, University of Nebraska-Lincoln, 343 E Hardin Hall, 3310 Holdrege Street, Lincoln, NE 68583-0961, USA

A R T I C L E I N F O

A B S T R A C T

Keywords: Antioxidant capacity Extraction Hazelnut shells HPLC-DAD HPLC-MS/MS Optimization Phenolics Response surface experiment

Hazelnut shells are the major byproduct of the hazelnut industry. The objectives of this study were to optimize the conditions for extracting phenolics and to identify and quantify the phenolics in hazelnut shells. Preliminary optimization showed that a high recovery of phenolics could be achieved with shell particle size less than 0.5 mm when extracted with acetone at 50 °C. Response surface experiments showed that a 10 g/l liquid to solid ratio, 58% acetone, and 12 h extraction time yielded the highest amount of phenolics. Twenty-seven phenolic compounds were identified in hazelnut shells by mass spectrometry. Coumaroylquinic acid, epicatechin gallate, quercetin, and six other phenolics were identified in hazelnut shells for the first time. The most abundant phenolics in hazelnut shells were catechin, epicatechin gallate, and gallic acid, as quantified by high performance liquid chromatography (HPLC). These results can be useful for the development of industrial extraction processes of natural antioxidants from hazelnut shells.

1. Introduction Hazelnuts (Corylus avellana L.) originated in the Mediterranean region and are now an important commercial crop in many countries. World production of hazelnuts achieved 488,110 metric tons (kernel basis) in 2015 and the production has increased by 50% in the past decade (INC, 2009, 2016). Turkey is the largest hazelnut producer in the world, producing 70% of the world’s total. The United States (U.S.A.) produced approximately 2% of the total (INC, 2016). In the U.S.A., 99% of hazelnuts are grown in the state of Oregon, and “Barcelona” is the dominant cultivar grown in the U.S.A. (Beyer, Grishina, Bardina, Grishin, & Sampson, 2002). Hazelnut kernels are rich in unsaturated fatty acids, essential amino acids, dietary fibres, vitamins, and minerals (Köksal, Artik, Şimşek, & Güneş, 2006). Due to their nutritious quality and unique flavour, they are widely used in dairy, bakery, coffee, spreads, confectionery products, and salads (Ozdemir & Akinci, 2004). Only 10% of hazelnuts are purchased as in-shell nuts and 90% of hazelnuts are used for industrial purposes as shelled nuts (Stévigny, Rolle, Valentini, & Zeppa, 2007). Hazelnut shells represent more than 50% of the total nut weight and they are the major byproduct in hazelnut industry production (Caglar & Aydinli,

2009). Hazelnut shells are composed of about 30% hemicelluloses, 27% celluloses, and 43% lignin, so they are mainly utilized as a low-value heat source (Demirbaş, 1999). Conversions of hazelnut shells into useful chemicals, such as methanol (Güllü & Demirbaş, 2001), hemicellulosic sugars (Arslan, Takaç, & Eken-Saraçoğlu, 2012), reducing sugar (Uzuner & Cekmecelioglu, 2014), and furfural (Demirbas, 2006) have been reported. Recently some efforts have been made to utilize hazelnut shells as a low cost raw material for phenolic compound extraction (Contini, Baccelloni, Massantini, & Anelli, 2008; Shahidi, Alasalvar, & LiyanaPathirana, 2007; Xu, Sismour, Parry, Hanna, & Li, 2012). Phenolic compounds are the primary bioactive components in plants. They have a wide range of health benefits, mainly due to their antioxidant properties, such as reactive oxygen species scavenging and inhibition, electrophile scavenging and metal chelation (Randhir, Lin, & Shetty, 2004). Phenolic compounds also exhibit pharmacological properties, such as anti-carcinogenic, anti-inflammatory, and anti-mutagenic effects, and anti-proliferative potential (Kaliora, Kogiannou, Kefalas, Papassideri, & Kalogeropoulos, 2014). Currently, many synthetic antioxidants are being used to retard the oxidation process, particularly in food systems. However, application of synthetic antioxidants in food products are of concern and are strictly regulated, due

Abbreviations: ANOVA, analyses of variance; Cs, concentration of solvent; DAD, diode array detector; DPPH, 2,2-diphenyl-1-picryl hydrazyl; DRSC, DPPH radical-scavenging capacity; FRAP, ferric reducing antioxidant power; GAE, gallic acid equivalents; HPLC, high performance liquid chromatography; MRM, multiple reaction monitoring; MS/MS, tandem mass spectrometry; m/z, mass-to-charge ratio; Ps, particle size; S/L, solid to liquid ratio; T, extraction temperature; t, extraction time; TE, trolox equivalents; TPC, total phenolic content ⁎ Corresponding author at: Department of Food Science and Technology, University of Nebraska-Lincoln, 1901 North 21st Street, Lincoln, NE 68588-6205, USA. E-mail addresses: [email protected] (B. Yuan), [email protected] (M. Lu), [email protected] (K.M. Eskridge), [email protected] (L.D. Isom), [email protected] (M.A. Hanna). http://dx.doi.org/10.1016/j.foodchem.2017.09.116 Received 30 May 2017; Received in revised form 14 August 2017; Accepted 21 September 2017 Available online 23 September 2017 0308-8146/ © 2017 Elsevier Ltd. All rights reserved.

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

speed (250 rpm) at a prescribed temperature. Afterwards, the extracts were centrifuged for 5 min at 5000g, and the residues were washed with 10 ml of distilled water and re-centrifuged twice under the same conditions. The supernatants were collected for phenolic content and antioxidant analyses. Each extraction was carried out in triplicate.

to potential health hazards (Park, Jung, Nam, Shahidi, & Kim, 2001). Consequently, the utilization of natural phenolic antioxidants, as alternatives, has raised considerable interest among food scientists, manufacturers and consumers. Although hazelnut shells are rich in phenolic compounds, very little is known about the extraction and composition of phenolic compounds from hazelnut shells. The objectives of this research were to investigate the optimum conditions for extracting phenolic compounds, using response surface experiments, and to characterize the phenolic composition in hazelnut shell extracts.

2.4. Experimental design 2.4.1. Preliminary study of the extraction conditions Before the optimization by response surface experiment, a first set of three tests was performed to identify the relevant extraction conditions used in the formal experiment, including the particle size of the hazelnut shells, the type of solvent, and the extraction temperature. First, a two-factorial design (4 levels of particle sizes × 4 levels of extraction time) was applied to evaluate the effects of particle size and extraction time on the extraction of phenolic antioxidants. The 4 levels of particle size were 1–2 mm, 0.5–1 mm, < 0.5 mm (Wiley mill), and < 0.5 mm (Wiley and burr mills). The 4 levels of extraction time were 2, 4, 6, and 12 h. Second, another two-factorial design (3 solvents × 3 levels of concentration) was used to evaluate the effects of solvents on the extraction of phenolics. The solvents used were methanol, ethanol, and acetone. Each solvent was tested at concentrations of 20%, 50%, and 80%. Lastly, a single-factorial design was used to evaluate the influence of temperature (30, 40, and 50 °C) on the extraction. All experiments were repeated at least twice for each treatment combination.

2. Materials and methods 2.1. Raw materials 2.1.1. Preparation of hazelnut shell powders “Lewis” cultivar hazelnuts were harvested and de-husked from Hazelnut Hill Farm in Corvallis, Oregon in the fall of 2012. “Lewis” is a hazelnut cultivar developed by Oregon State University in 1997. “Lewis” has a higher yield efficiency and a smaller tree size than “Barcelona” cultivar (Mehlenbacher, Azarenko, Smith, & McCluskey, 2000). Hazelnut shells were first ground to pass through a 10-mesh (2 mm) sieve, using a Wiley mill (Standard Model No. 3, Arthur H. Thomas Co., U.S.A.). The first half of the screened materials was further ground in a burr mill (Mil-Rite grain mill, Retsel Corp., U.S.A.) to pass through a 35mesh (0.5 mm) sieve. The particle size of hazelnut shells grinding, by both Wiley and burr mills, was considered as smaller than 0.5 mm. The second half of the screened materials was sifted to pass through 18mesh (1.0 mm) and 35-mesh (0.5 mm) sieves, sequentially. By grinding with a Wiley mill and sifting, the particle sizes of the 3 fractions of hazelnut shells were 1–2 mm, 0.5–1 mm, and < 0.5 mm. The weights of different fractions of hazelnut shells were recorded. The ground samples were stored at -20 °C for further analysis.

2.4.2. Response surface experiment A full factorial experimental design (3 × 3 × 5) was used to determine the optimum extraction condition for phenolic antioxidants. The variables were solid to liquid ratio (S/L = 10, 30, and 50 g/l), concentration of solvent (Cs = 20, 50, and 80%), and extraction time (t = 1, 2, 4, 6, and 12 h). The particle size (< 0.5 mm), type of solvent (acetone) and extraction temperature (50 °C) were kept constant. All of the experiments were repeated at least twice for each treatment combination in completely randomized designs.

2.1.2. Proximate composition of hazelnut shells Proximate composition, including moisture, ash, protein, fat, and carbohydrate contents was determined on hazelnut shell powders. Moisture, ash, protein and fat contents were analyzed, following the standard methods (AOAC, 2000). Crude fat was determined by using an extraction unit (HT1043, Soxtec, U.S.A.) with hexane as the extraction solvent. Crude protein was analyzed with a nitrogen analyzer unit (Leco FP-528, Leco Corporation 3000, U.S.A.) with 6.25 as nitrogen to protein conversion factor. Carbohydrate content was calculated by subtracting contents of other compositions from 100%. Analyses were performed in triplicate. Data were reported as a percentage of the wet weight of the hazelnut shell powder.

2.5. Analyses of the response variables 2.5.1. Total phenolic content (TPC) The TPC was measured according to the method of Siriwardhana and Shahidi (2002) with modification. Equal volumes (0.1 ml) of FolinCiocalteau reagent and diluted extract were mixed, then 1.0 ml of sodium carbonate solution (75 g/l) was added to the content. After 1 h of incubation at room temperature in the dark, 200 µl of the mixture were transferred into the designated well of a 96-well microplate. The absorbance was read at 725 nm, using a microplate reader (BioTek Instruments, U.S.A.). Gallic acid standard solutions were used for calibration. The results of TPC were expressed as mg gallic acid equivalents (GAE) /g of shell.

2.2. Chemicals Folin-Ciocalteu reagent, 2,4,6-tri (2-pyridyl)-1,3,5-triazine (TPTZ), 2,2-diphenyl-1-picryl hydrazyl (DPPH), trolox, ferulic acid, gallic acid, coumaric acid, 3-O-caffeoylquinic acid, taxifolin, catechin and epicatechin were purchased from Sigma-Aldrich (U.S.A.). Protocatechuic acid, phlorizin, quercitrin, sodium carbonate, ferric chloride hexahydrate, acetone, ethanol, and methanol were purchased from Fisher Scientific (U.S.A.). Quercetin, kaempferol, and epicatechin gallate were obtained from Cayman Chemical (U.S.A.). Myricitrin was purchased from VWR (U.S.A.).

2.5.2. Antioxidant capacity 2.5.2.1. DPPH radical-scavenging capacity (DRSC). The DRSC values of antioxidants in the hazelnut shell extracts were evaluated, based on the method by Lu, Yuan, Zeng, and Chen (2011). The DPPH radical solution was prepared by dissolving 3.5 mg of DPPH radical in 100 ml of ethanol. Accurately, 3 ml of DPPH radical ethanolic solution were added to 0.15 ml of properly diluted hazelnut shell extract. The mixture was shaken vigorously for 1 min and left to stand at room temperature in the dark for 30 min. Absorbance was measured against the blank reagent at 517 nm (Evolution 201 UV–Visible spectrophotometer, Thermo Fisher, China). All determinations were carried out in triplicate. The percent inhibition of DPPH radical was calculated according to the equation as shown below

2.3. Shaking bath extraction The shaking bath extraction was carried out in a temperature-controlled incubator shaker (Innova 26, New Brunswick Scientific, U.S.A.). Each sample was macerated with 10 ml of extraction solvent in a closed 50 ml centrifuge tube. The centrifuge tubes were shaken at a constant

Inhibition of DPPH radical (%) = [(Ac −As )/ Ac ] × 100

(1)

where Ac is the absorbance of the control solution, and As is the 8

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

acetic acid in water. Two different gradient elution methods were applied, due to the difficulty in separating the phenolic compounds, simultaneously. For the separation of myricitrin and epicatechin gallate, the following elution method was applied: linear gradient elution from 5 to 22% A for 0–10 min, 22 to 24% A for 10–20 min, 24 to 95% A for 20–25 min, 95 to 5% A for 25–30 min; and then isocratic elution of 5% A for the next 10 min. For the separation of other phenolic compounds (gallic acid, coumaric acid, protocatechuic acid, taxifolin, quercetin, quercitrin, catechin, epicatechin, 3-O-caffeoylquinic acid, and phlorizin), a second gradient elution method was used. Linear gradient elution from 5 to 22% A for 0–25 min, 22 to 23% A for 25–37 min, 23 to 35% A for 37–50 min, 35 to 40% A for 50–60 min, 40 to 95% A for 60–65 min, 95 to 5% A for 65 to 67 min; then isocratic elution of 5% A, 67–75 min. The column temperature was ambient temperature, the flow rate was 0.8 ml/min, and the injection volume was 10 µl. Separated phenolic compounds were monitored at wavelengths of 254, 280, 320, and 370 nm to enhance the accuracy of quantification. Quantifications of phenolics were done by comparing the retention times and UV spectra to the individual external standards. The dimer of prodelphinidin, dimer and trimer of procyanidin, pentose ester of coumaric acid, hexose esters of syringic acid, galloylquinic, and coumaroylquinic acid were not quantified by HPLC-DAD, mostly due to the commercial unavailability of phenolic standards.

absorbance of the test solution. Trolox standard solutions were used for calibration The DRSC values were expressed as μmol trolox equivalents (TE)/g shell. 2.5.2.2. Ferric reducing antioxidant power (FRAP). The FRAP assay was conducted according to the method of Benzie and Strain (1996) with modification. The FRAP reagent included 10 mM TPTZ solution in 40 mM HCl, 20 mM FeCl3 solution and 0.3 M acetate buffer (pH 3.6) in proportions of 1:1:10 (v/v). Each diluted extract (50 µl) was mixed with 3 ml of freshly prepared FRAP reagent and the reaction mixtures were incubated at 37 °C for 30 min. Then, 200 µl of the mixture were transferred into the microplate reader and the absorbance was read at 593 nm. Trolox standard solutions were used for calibration. The FRAP values were expressed as μmol TE/g of shell. 2.6. Phenolic composition analysis 2.6.1. Preparation of concentrated extracts To determine the phenolic composition, the hazelnut shells were extracted under the obtained optimal extraction conditions. The extracts were concentrated and analyzed by high performance liquid chromatography (HPLC) with a diode array detector (DAD) and HPLC tandem mass spectrometry (MS/MS). The extracts were dried, using a Büchi Rotavapor (R-144, Büchi Corp., U.S.A.) at 40 °C, and the residues were dissolved in 2 ml of 50% aqueous methanol. For the hydrolysis of the phenolic extracts, organic solvents from the extracts were removed under vacuum at 40 °C with a Büchi Rotavapor. The aqueous residue of the extract and the shell residue from soluble phenolic compounds extraction were frozen at −40 °C and then lyophilized in a Lyph-Lock 12 freeze-dryer (Labconco, U.S.A.).

2.7. Statistical analyses TPC values from the response surface experiment (3 × 3 × 5 factorial) were analyzed with analyses of variance (ANOVA) for a third order polynomial model (Table 2). The quality of the fit of the polynomial model was expressed by the coefficient of determination R2. The statistical significance was checked, using an F-test of the entire model and lack of fit was checked using an ANOVA table. For all statistical analysis, P < 0.05 was considered as statistical significance. The error bars in all Figures correspond to the standard deviations. All statistical analyses were performed with the software SAS version 9.4 (SAS Institute Inc., U.S.A.).

2.6.2. Identification of phenolic compounds by HPLC-MS/MS analysis Phenolic compounds in hazelnut shell extracts were identified and then qualitatively analyzed, using a Waters 2695 separation module and Quattro-Micro triple-quadrupole mass spectrometer equipped with an electrospray ionization source in negative ion mode (Waters, U.S.A.). Chromatographic separations were performed, using a HyPURITY C18 HPLC column (250 mm × 2.1 mm ID, 5 mm particle size) at a temperature of 50 °C and a flow rate of 0.2 ml/min. A gradient elution of mobile phase A (1% v/v formic acid in acetonitrile) and mobile phase B (1% v/v formic acid in distilled de-ionized H2O) was used for the chromatographic separation. The gradient profile was set as follows: initial conditions at 0% A holding for 5 min, followed by linear gradient to reach 40% A at 30 min, solvent A immediately increased to 100% and held for 5 min, then immediately reduced back to initial conditions (0% A), and held for 5 min. Multiple reaction monitoring (MRM) experiments were performed on the basis of literature for the tentative identification of phenolic compounds (Arráez-Román, Fu, Sawalha, SeguraCarretero, & Fernández-Gutiérrez, 2010; Ciemniewska-Żytkiewicz et al., 2015; Del Rio, Calani, Dall’Asta, & Brighenti, 2011; Hofmann, Nebehaj, & Albert, 2015; Ma et al., 2014; Weisz, Kammerer, & Carle, 2009). Mass spectrometry analysis was performed under the following conditions: collision gas (Ar, 4.0 × 10−3 Torr); desolvation gas (N2, 750 l/h); cone gas (N2, 50 l/h); desolvation temperature (350 °C); source temperature (120 °C); capillary voltage (3 kV); extractor voltage (2 V). The MRM channels, cone voltages and collision energies used for each analyte are shown in Table 1.

3. Results and discussion 3.1. Hazelnut shell composition The Lewis cultivar hazelnut shells used in the study contained 7.24% moisture, 1.91% protein, 1.44% crude fat, 1.21% ash, 88.2% carbohydrate (% weight of the shell on a wet basis). These results agreed well with the study of Xu, Sismour, et al. (2012). Carbohydrates were the major components in hazelnut shells, which were mainly composed of fibres. 3.2. Preliminary optimization of particle sizes, solvents, and temperature 3.2.1. Effect of particle sizes Comparing the TPC and the antioxidant capacity values of the three fractions of hazelnut shells obtained by grinding with Wiley mill and sifting, Fig. 1 shows that TPC and antioxidant capacity (FRAP and DRSC) of the extracts significantly increased with the reduction of the particle size of hazelnut shells among all the extraction times (2–12 h). The TPC and antioxidant capacity of the extracts from ground shells smaller than 0.5 mm were about double the values of the ground shells with particle size of 0.5–1.0 mm. However, the yields of the different fractions of hazelnut shells, determined by weight after grinding by Wiley mill and sifting, were: 19.4% of particle size smaller than 0.5 mm, 52.0% of particle size between 0.5 and 1.0 mm, and 28.6% of particle size between 1.0 and 2.0 mm. This means that a lower amount of phenolic compounds, with lower antioxidant capacity, would be obtained if the hazelnut shells ground only by Wiley mill were used directly for phenolic extraction. Therefore, further particle size

2.6.3. Quantitative analysis of phenolic compounds by HPLC-DAD Because there was no ultra-violet detector equipped with the HPLCMS/MS, an Agilent 1260 HPLC-DAD system (Agilent Technologies, U.S.A.) was used for the quantitative analyses of phenolic compounds. A Shodex C18 column (250 mm × 4.6 mm inner diameter, 5 µm particle size) was employed for the separation of phenolic compounds. The mobile phases were (A) 0.5% acetic acid in acetonitrile and (B) 2.0% 9

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

Table 1 Tentative identification of phenolic compounds in hazelnut shells by HPLC-MS/MS in MRM mode. Phenolic compounds

Ferulic acid Gallic acid Vanillic acid Coumaric acid Protocatechuic acid Sinapic acid Caffeoylquinic acid Galloylquinic acid Coumaroylquinic acid* Feruloylquinic acid Pentose ester of coumaric acid* Hexose ester of syringic acid* Catechin/Epicatechin Epicatechin gallate* Taxifolin Quercetin* B-type dimer of PD2* B-type dimer of PC3* B-type trimer of PD B-type trimer of PC* B-type dimer gallate Kaempferol rhamnoside* Myricitrin (Myricetin-3-O-rhamnoside) Quercitrin (Quercetin-3-O-rhamnoside) Quercetin-pentoside Phlorizin (Phloretin-2-O-glucoside) Myricetin Isohamnetin rutinoside Rutin (Quercetin-3-O-rutinoside)

Experimental Parameters

Results

MRM (m/z)

Cone voltage (V)

Collision energy (eV)

Signal intensity1

Retention time (min)

193 → 134 169 → 125 167 → 123 163 → 119 153 → 109 223 → 164 353 → 191 343 → 191 337 → 191 367 → 191 295 → 163 359 → 197 289 → 137 441 → 169 303 → 125 301 → 151 593 → 289 577 → 289 881 → 125 865 → 125 729 → 289 431 → 285 463 → 317 447 → 301 433 → 301 435 → 273 317 → 151 623 → 315 609 → 301

30 25 30 30 25 30 30 30 30 30 25 25 35 35 30 35 35 35 35 35 35 35 35 35 30 35 35 35 35

30 20 30 30 20 30 30 30 30 30 20 20 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30

– 3.66E + 04 – 1.68E + 03 3.71E + 04 – 3.58E + 03 1.08E + 03 2.09E + 04 – 1.14E + 03 2.00E + 04 1.06E + 04 1.18E + 03 3.16E + 03 1.12E + 04 1.62E + 04 5.41E + 03 – 1.10E + 03 – 7.06E + 04 1.58E + 04 6.16E + 04 – 7.90E + 03 – – –

– 4.03 – 16.53, 21.48 5.89 – 6.57, 6.82, 18.51 34.71 16.53 – 17.70, 20.55 18.63 15.35, 15.66 23.27 23.52 32.73 4.34 8.74, 12.08, 19.37, 21.23 – 19.87 – 27.66 23.58 25.87 – 27.41 – – –

– Not detected. 1 Phenolic compounds with signal intensity lower than 1000 were not shown. 2 PD = prodelphinidin. 3 PC = procyanidin. * Phenolic compounds identified in hazelnut shells for the first time.

reduction should be applied for higher yield of phenolic compounds with higher antioxidant capacities. Comparing the two fractions with particle size smaller than 0.5 mm,

Fig. 1 also shows that the TPC and antioxidant capacity of the hazelnut shells further ground by burr mill were just slightly lower than the values of extracts ground by Wiley mill only. In addition, the yield of

Table 2 Parameter estimates and fitness of the models without or with transformation of time for the response variable of TPC. Transformation of time (t)

Lack of fit Parameter estimates

R2 Adjusted R2 1 2 3

Intercept Ratio Ratio * Ratio Conc1 Ratio * Conc Ratio * Ratio * Conc Conc * Conc Ratio * Conc * Conc tn2 Conc * tn Conc * Conc * tn tn * tn Ratio * tn Ratio * Ratio * tn Ratio * * tn * tn Conc * Conc * tn Ratio * Conc * tn

Model 1

Model 2

Model 3

Untransformed

t1 = ln (t)

t2 = t^0.5

0.0295 6.444799 −0.056496 0.00023 0.091654 0.00211 −9.29E−06 −0.000825 −1.44E−05 0.396950 0.005242 −0.000044 −0.022946 −0.005261 0.000050 0.000157 ns3 ns 0.9761 0.9733

0.1701 7.259397 −0.076481 0.00048 0.083583 0.00211 −9.29E−06 −0.000753 −1.44E−05 0.104737 0.026966 −0.000229 0.048604 ns ns ns ns ns 0.9787 0.9768

0.1034 5.6239586 −0.0382495 7.797E−05 0.0646512 0.00211 −9.29E−06 −0.0005966 −1.44E−05 1.5101634 0.0257612 −0.0002164 −0.2472337 −0.0251435 0.0001944 0.0027409 ns ns 0.9795 0.9771

Conc = concentration. tn = t for time untransformed model (Model 1); n = 1 and 2 corresponded to two time transformed models (Model 2 and 3, respectively). ns = item was not significant in Type I test.

10

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

Fig. 1. Effect of particle size on (A) TPC, (B) FRAP, and (C) DRSC when T = 50 °C, Cs = 50%, and S/L = 30 g/l. Bars marked with different letters are significantly different (p < 0.05) for individual responses.

particles and also increases the contact area of the extracting solvent with the matrix, which contributes to the enhanced extraction (Sun, Liu, Chen, Ye, & Yu, 2011). The results may also be explained by the distribution of phenolic compounds in different parts of hazelnut shells (Romdhane & Gourdon, 2002). The inner part of the shells are easy to crush, thereby forming the majority of the group with particles smaller than 0.5 mm. The inner parts of the shells are similar in appearance to hazelnut skins, and hazelnut skins have higher phenolic content and antioxidant potential than have hazelnut shells (Contini et al., 2008;

hazelnut shells with particle size smaller than 0.5 mm by grinding with burr mill after Wiley mill was over 95%. Thus, the hazelnut shells with particle size smaller than 0.5 mm, ground by Wiley mill and then burr mill, were used for the rest of the tests. The extended extraction time did not significantly improve the extraction efficiency compared to the reduction of particle size. Thus, smaller particle size could be applied for shortening the extraction time and increasing the yield of antioxidant phenolics. Reducing particle size increases the specific surface area of the shell

11

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

12.00 l TPC (mg GAE/g)

10.00 a 8.00

b

m

m

c

3.3.2. Selection of model for TPC prediction Normally, second order polynomial models have been employed for optimization of shaking bath extractions of TPC (Radojkovic et al., 2012; Stévigny et al., 2007; Xu, Bao, Gao, Zhou, & Wang, 2012; Yim et al., 2012). However, models with the three factors were not valid for this study because the lack of fit was significant when time was used directly in the second order model. As alternatives, several transformations of the variable “time” were carried out, as shown in Table 2. The parameter estimates of new models with their probability of lack of fit, coefficient of determination (R2), and adjusted coefficient of determination (adjusted R2) also are listed in Table 2. Both predicted models with transformation of time had satisfactory goodness of fit (p > 0.05 for lack of fit) and coefficient of determination (R2 = 0.98), so theoretically, either of these models could be used to represent the influences of the factors on the extraction. Compared with model 3, model 2 had a larger p value for lack of fit, and it had fewer terms; that made it a simpler model. As a result, the model 2 was selected and expressed as below:

x

y z

6.00

Methanol Ethanol

4.00

Acetone 2.00 0.00 20

50 Solvent concentration (%)

80

Fig. 2. Effect of type of solvent and its concentration on TPC when Ps < 0.5 mm, T = 50 °C, t = 2 h, and S/L = 30 g/l. Bars marked with different letters are significantly different within each concentration (p < 0.05).

Shahidi et al., 2007). On the other hand, the outer parts of the shells are more resistant to crushing, thereby forming a class with large particles; they are therefore less extractable.

Y = 7.26−7.65 × 10−2X1 + 4.80 × 10−4X12 + 8.36 × 10−2X2 + 2.11 × 10−3X1 X2 −9.29 × 10−6X12 X2 −7.53 × 10−4X22 −1.44 × 10−5X1 X22 + 1.05 × 10−1X3 + 2.70 × 10−2X2 X3−2.29 × 10−4X22 X3 + 4.86 × 10−2X32

3.2.2. Effect of different solvents Many different solvents have been utilized for phenolic compounds extraction. Aqueous methanol (Ciarmiello et al., 2014), ethanol (Shahidi et al., 2007), and acetone (Stévigny et al., 2007) are the most commonly used solvents for phenolic extraction from tree nut shells. Thus, these three aqueous solvents were chosen for phenolics extraction from hazelnut shells in this study. At each solvent concentration point, aqueous acetone solution extracted significantly higher amounts of phenolics than did either methanol or ethanol (Fig. 2). A large portion of phenolic compounds in hazelnut shells is tannins (Contini et al., 2008; Xu, Sismour, et al., 2012) and acetone is considered a good solvent for tannins extraction (Hagerman, Harvey-Mueller, & Makkar, 2000). Fig. 2 also shows that higher amounts of TPC were extracted by 50% aqueous solution compared with 20% and 80% aqueous solutions for all three solvents. Aqueous acetone had the best extracting capacity and was used as the extraction solvent in subsequent experiments.

where Y is the predicted TPC, X1 is solid to liquid ratio, X2 is acetone concentration, and X3 is natural logarithm of extraction time. 3.3.3. Effect of extraction conditions on TPC All three factors had linear and quadratic main effects on phenolic content (P < 0.05), and there were several interaction effects, but no significant three-way interaction effects were found. The relationships between independent and dependent variables are illustrated by the response surface plots shown in Fig. 3. Acetone concentration had a quadratic effect on the phenolic content, which increased with increasing acetone concentration from 20% to about 60%, and decreased thereafter (Fig. 3A and B). Phenolic content increased with the decrease in solid to liquid ratio at low acetone concentration, but the trend became weak at higher concentration (Fig. 3C–E). It was observed that the quadratic effect of the ratio was not obvious although its second order item ( X12 ) existed in the model. Similarly, the effect of natural logarithm of time (X3) was close to be linear on phenolic content (Fig. 3A and C), but the change of TPC with time was logarithmic (Fig. 3B and D). Phenolic content increased dramatically with the increase in extraction time in the first several hours, but then the rate increase was reduced. Based on the surface plots, maximum TPC was obtained at the highest time point, which was confirmed by the optimal extraction conditions obtained, based on the predicted model: solid to liquid ratio of 10 g/l, acetone concentration of 58%, and extraction time of 12 h for the maximum TPC of 12.1 mg GAE/g of shell. The TPC of the extract obtained under the optimal conditions was 12.0 ± 0.15 mg GAE/g of shell, which was very close to the value predicted by the model with the coefficient of variation of 1.16%. This validates the optimal conditions. Besides the work done in this study, only Stévigny et al. (2007) have reported the optimal extraction conditions of phenolics in hazelnut shells. The maximum phenolic content was predicted as 6.67 mg GAE/g of shell under extraction conditions of 55.7% ethanol and 108.7 min of extraction time. The maximum phenolic content predicted in their study was only about half of the content obtained in this study. All the extractions in their study were conducted at room temperature, while the extraction temperature in this study was 50 °C. The results of the present study show that aqueous acetone was a better solvent than aqueous ethanol for phenolic extraction from hazelnut shells, and a higher extraction temperature could accelerate the extraction process. For the first time, solid to liquid ratio was considered in the optimal extraction of phenolics in hazelnut shells. The importance of the influence of solid to liquid ratio on phenolic compounds extraction had

3.2.3. Effect of temperature Extraction temperature is one of the most important factors contributing to increased recoveries of solute from solid matrices. Since the boiling point of acetone is 56 °C, at atmospheric pressure, higher temperature was not studied, due to the limitation of the experimental equipment. Our results showed that TPC showed an increasing trend with increasing temperature. The results of TPC were 8.01, 8.79, and 10.5 mg GAE/g of shell at 30, 40, and 50 °C, respectively. Higher temperature increased the solubility, as well as the diffusivity of phenolic compounds (Sun et al., 2011). Moreover, higher temperature can cause an increase in cellular pressure which may cause cell rupture and opening of the cell matrix, resulting in increased availability of phenolic compounds to be extracted by the solvent. Additionally, surface tension and solvent viscosity decrease with increased temperature, which improves sample wetting and matrix penetration, respectively (Khajeh, Moghaddam, & Sanchooli, 2010; Sparr Eskilsson & Björklund, 2000).

3.3. Effect of solid to liquid ratio, acetone concentration, and extraction time 3.3.1. General Effects of solid to liquid ratio (10, 30, 50 mg/ml), acetone concentration (20, 50, 80%), and extraction time (1, 2, 4, 6, 12 h) on total phenolic extraction were studied. 12

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

Fig. 3. Surface plot of the TPC of hazelnut shell extract as affected by solid to liquid ratio, acetone concentration, and extraction time (B, D, and E) or natural logarithm of extraction time (A and C).

compound identification. Among these reported methods, the HPLCMS/MS method, in MRM mode, is a powerful approach for simultaneous determination of multiple components, based on the mass-tocharge ratio (m/z) of the molecular ion ([M−H]−) and its corresponding characteristic daughter ion for individual phenolic compounds. In this study, the HPLC-MS/MS was used to tentatively identify the phenolic compounds from the hazelnut shell extract and the results are summarized in Table 1. A minimum signal intensity threshold of 1000 was applied in the HPLC-MS/MS instrument because compounds with signal intensity lower than 1000 are considered to be masked by background noise signals. Some of these compounds had several

been proven by many studies (Maran, Priya, & Nivetha, 2015; Purohit & Gogate, 2015; Radojkovic et al., 2012). The influence of solid to liquid ratio relates to mass transfer principles. Concentration gradient between the solid and the bulk of the liquid is the driving force of mass transfer, which is greater when a lower solid to liquid ratio is used. 3.4. Identification of phenolic compounds by HPLC-MS/MS analysis Many analytical techniques, such as UV spectrometry, GC–MS, HPLC-DAD, and HPLC-MS/MS, have been applied widely to phenolic 13

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

Fig. 4. HPLC-DAD Chromatogram of the standard mixture of 13 phenolic compounds at 280 nm and the quantification results of phenolic compounds in hazelnut shells. (A) First gradient elution separated two phenolic compounds; (B) Second gradient elution separated eleven phenolic compounds; (C) The identities of the peaks and the amount of the detected phenolic compounds in hazelnut shells are shown in part C.

3.5. Quantitative analysis of the identified individual phenolic compounds by HPLC-DAD

different retention times under the same name, which means that these compounds had isomers. For example, catechin and epicatechin are isomers with the same molecular ion [M−H]− at m/z 289 and characteristic fragment ion at m/z 137. Similarly, a B-type procyanidin dimer can be formed from two procyanidin molecules with C4 → C8 or C4 → C6 bonds in the alpha or beta position. Four isomers of B-type procyanidin were identified tentatively in the Lewis cultivar hazelnut shells. In total twenty-seven phenolic compounds were recognized tentatively in the hazelnut shell extracts. Four phenolic acids (gallic acid, protocatechuic acid, and two coumaric acids), eight esters of phenolic acids (3 caffeoylquinic acids, galloylquinic acid, coumaroylquinic acid, 2 pentose esters of coumaric acid, and hexose ester of syringic acid), eleven flavonoids (catechin, epicatechin, epicatechin gallate, taxifolin, quercetin, one prodelphinidin dimer, four procyanidin dimers, and one procyanidin trimer), and four flavonoid glycosides (kaempferol rhamnoside, myricitrin, quercitrin, and phlorizin) were detected. Limited phenolic compounds from hazelnut shells were identified (Ciemniewska-Żytkiewicz et al., 2015; Nazzaro et al., 2012; Shahidi et al., 2007). In this study, many more phenolic compounds were identified tentatively. Coumaroylquinic acid, pentose ester of coumaric acid, hexose ester of syringic acid, epicatechin gallate, quercetin, dimer and trimer of procyanidin, dimer of prodelphinidin, and kaempferol rhamnoside were identified in the hazelnut shells for the first time. Furthermore, this is the first-ever detection of coumaroylquinic acid in hazelnuts (including kernels, skins, husks), which also has been reported in sunflower hulls (Weisz et al., 2009).

Based on the result of HPLC-MS/MS analyses, quantifications of the phenolic compounds using pure phenolic standards by HPLC-DAD, were conducted. Typical HPLC-DAD chromatograms of the standard mixture of 13 phenolic compounds at 280 nm are shown in Fig. 4, in which part A demonstrates the first gradient elution of myricitrin and epicatechin gallate and part B demonstrates the second gradient elution of eleven phenolic compounds. The identified and quantified phenolic compounds in the extract of Lewis cultivar hazelnut shells are shown in Fig. 4(C). Seven phenolic compounds were quantified, two of which were phenolic acids (gallic acid and protocatechuic acid), three of which were flavanols (catechin, epicatechin, and epicatechin gallate), and two of which were flavonol (quercetin) and its glycoside (quercitrin). Among the quantified phenolic compounds, catechin was the most abundant one in the extract of hazelnut shells with the concentration of 176 µg/g of shells, followed by epicatechin gallate, gallic acid, quercetin, protocatechuic acid, quercitrin, and epicatechin. Gallic acid was identified as the most abundant phenolic compound in hazelnut shell by two other studies, with 80% ethanol solution as the extraction solvent (Ciemniewska-Żytkiewicz et al., 2015; Shahidi et al., 2007). In this study, 58% acetone solution was determined as the optimum condition by response surface experiment and then utilized for extraction for the purpose of identification and quantification of phenolic compounds. Nazzaro et al., 2012 identified and quantified several phenolic compounds in extracts of chestnut and hazelnut shells after different solvent systems were used for extraction. They found differences in the phenolic composition when different solvents were used. Therefore, the extraction conditions not only affect the TPC and antioxidant capacity, but also have great influence on the phenolic composition.

14

Food Chemistry 244 (2018) 7–15

B. Yuan et al.

production of briquettes using pyrolytic oil. Energy, 24(2), 141–150. Demirbas, A. (2006). Furfural production from fruit shells by acid-catalyzed hydrolysis. Energy Sources, Part A, 28(2), 157–165. Güllü, D., & Demirbaş, A. (2001). Biomass to methanol via pyrolysis process. Energy Conversion and Management, 42(11), 1349–1356. Hagerman, A., Harvey-Mueller, I., & Makkar, H. (2000). Quantification of tannins in tree foliage–a laboratory manual. Vienna, Austria. Hofmann, T., Nebehaj, E., & Albert, L. (2015). The high-performance liquid chromatography/multistage electrospray mass spectrometric investigation and extraction optimization of beech (Fagus sylvatica L.) bark polyphenols. Journal of Chromatography A, 1393, 96–105. International Nuts and Dried Fruits Foundation, INC (2009). Nuts and dried fruits, global statistical review 2004–2009. International Nuts and Dried Fruits Foundation, INC (2016). Nuts & dried fruits global statistical review 2015/2016. Kaliora, A. C., Kogiannou, D. A., Kefalas, P., Papassideri, I. S., & Kalogeropoulos, N. (2014). Phenolic profiles and antioxidant and anticarcinogenic activities of Greek herbal infusions; balancing delight and chemoprevention? Food Chemistry, 142, 233–241. Khajeh, M., Moghaddam, A. R. A., & Sanchooli, E. (2010). Application of Doehlert design in the optimization of microwave-assisted extraction for determination of zinc and copper in cereal samples using FAAS. Food Analytical Methods, 3(3), 133–137. Köksal, A.İ., Artik, N., Şimşek, A., & Güneş, N. (2006). Nutrient composition of hazelnut (Corylus avellana L.) varieties cultivated in Turkey. Food Chemistry, 99(3), 509–515. Lu, M., Yuan, B., Zeng, M., & Chen, J. (2011). Antioxidant capacity and major phenolic compounds of spices commonly consumed in China. Food Research International, 44(2), 530–536. Ma, Y., Kosińska-Cagnazzo, A., Kerr, W. L., Amarowicz, R., Swanson, R. B., & Pegg, R. B. (2014). Separation and characterization of soluble esterified and glycoside-bound phenolic compounds in dry-blanched peanut skins by liquid chromatography-electrospray ionization mass spectrometry. Journal of Agricultural and Food Chemistry, 62(47), 11488–11504. Maran, J. P., Priya, B., & Nivetha, C. V. (2015). Optimization of ultrasound-assisted extraction of natural pigments from Bougainvillea glabra flowers. Industrial Crops and Products, 63, 182–189. Mehlenbacher, S. A., Azarenko, A. N., Smith, D. C., & McCluskey, R. (2000). Lewis' hazelnut. HortScience, 35(2), 314–315. Nazzaro, M., Mottola, M. V., La Cara, F., Del Monaco, G., Aquino, R. P., & Volpe, M. G. (2012). Extraction and characterization of biomolecules from agricultural wastes. Chemical Engineering, 27. Ozdemir, F., & Akinci, I. (2004). Physical and nutritional properties of four major commercial Turkish hazelnut varieties. Journal of Food Engineering, 63(3), 341–347. Park, P.-J., Jung, W.-K., Nam, K.-S., Shahidi, F., & Kim, S.-K. (2001). Purification and characterization of antioxidative peptides from protein hydrolysate of lecithin-free egg yolk. Journal of the American Oil Chemists' Society, 78(6), 651–656. Purohit, A. J., & Gogate, P. R. (2015). Ultrasound-assisted extraction of β-carotene from waste carrot residue: Effect of operating parameters and type of ultrasonic irradiation. Separation Science and Technology null–null. Radojkovic, M., Zekovic, Z., Jokic, S., Vidovic, S., Lepojevic, Z., & Milosevic, S. (2012). Optimization of solid-liquid extraction of antioxidants from black mulberry leaves by response surface methodology. Food Technology and Biotechnology, 50(2), 167. Randhir, R., Lin, Y. T., & Shetty, K. (2004). Phenolics, their antioxidant and antimicrobial activity in dark germinated fenugreek sprouts in response to peptide and phytochemical elicitors. Asian Pacific Journal of Clinical Nutrition, 13(3), 295–307. Romdhane, M., & Gourdon, C. (2002). Investigation in solid–liquid extraction: Influence of ultrasound. Chemical Engineering Journal, 87(1), 11–19. Shahidi, F., Alasalvar, C., & Liyana-Pathirana, C. M. (2007). Antioxidant phytochemicals in hazelnut kernel (Corylus avellana L.) and hazelnut byproducts. Journal of Agricultural and Food Chemistry, 55(4), 1212–1220. Siriwardhana, S. S. K. W., & Shahidi, F. (2002). Antiradical activity of extracts of almond and its by-products. Journal of the American Oil Chemists' Society, 79(9), 903–908. Sparr Eskilsson, C., & Björklund, E. (2000). Analytical-scale microwave-assisted extraction. Journal of Chromatography A, 902(1), 227–250. Stévigny, C., Rolle, L., Valentini, N., & Zeppa, G. (2007). Optimization of extraction of phenolic content from hazelnut shell using response surface methodology. Journal of the Science of Food and Agriculture, 87(15), 2817–2822. Sun, Y., Liu, D., Chen, J., Ye, X., & Yu, D. (2011). Effects of different factors of ultrasound treatment on the extraction yield of the all-trans-β-carotene from citrus peels. Ultrasonics Sonochemistry, 18(1), 243–249. Uzuner, S., & Cekmecelioglu, D. (2014). Hydrolysis of hazelnut shells as a carbon source for bioprocessing applications and fermentation. International Journal of Food Engineering, 10(4), 799–808. Weisz, G. M., Kammerer, D. R., & Carle, R. (2009). Identification and quantification of phenolic compounds from sunflower (Helianthus annuus L.) kernels and shells by HPLC-DAD/ESI-MSn. Food Chemistry, 115(2), 758–765. Xu, P., Bao, J., Gao, J., Zhou, T., & Wang, Y. (2012). Optimization of extraction of phenolic antioxidants from Tea (Camellia Sinensis L.) fruit peel biomass using response surface methodology. BioResources, 7(2), 2431–2443. Xu, Y., Sismour, E. N., Parry, J., Hanna, M. A., & Li, H. (2012). Nutritional composition and antioxidant activity in hazelnut shells from US-grown cultivars. International Journal of Food Science & Technology, 47(5), 940–946. Yim, H. S., Chye, F. Y., Koo, S. M., Matanjun, P., How, S. E., & Ho, C. W. (2012). Optimization of extraction time and temperature for antioxidant activity of edible wild mushroom, Pleurotus porrigens. Food and Bioproducts Processing, 90(2), 235–242.

4. Conclusions The choice of extraction conditions is of importance for the recovery of high contents of phenolic compounds. In this study, a particle size smaller than 0.5 mm, aqueous acetone solution, and extraction temperature of 50 °C were selected for high recovery of total phenolics. Response surface experiments were then successfully employed to optimize the extraction conditions of phenolic compounds in the extracts of Lewis hazelnut shells. The optimum conditions predicted were extracting with 58% aqueous acetone solution and using a solid to liquid ratio of 10 g/l for 12 h. Under the optimal conditions, the experimental value was in agreement with the predicted value, which yielded the highest TPC of 12.0 ± 0.15 mg GAE/g of shell. LC-MS/MS was used to quantify the phenolic compounds in Lewis hazelnut shells. Several phenolic compounds were identified in the hazelnut shells for the first time, including coumaroylquinic acid, epicatechin gallate, quercetin, kaempferol rhamnoside, pentose ester of coumaric acid, hexose ester of syringic acid, dimer and trimer of procyanidin, and dimer of prodelphinidin. Quantification of the phenolic compounds in the extract from Lewis hazelnut shells showed that catechin, epicatechin gallate, and gallic acid were the most abundant phenolic acids. This study can be useful to the development of industrial extraction processes of phenolic compounds from hazelnut shells. The hazelnut shells are rich in phenolic compounds and are high in antioxidant capacity, suggesting that hazelnut shells could be an innovative source for extracting natural antioxidants for use in the food industry. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgment The authors wish to thank Dr. Devin Rose, Dr. Junyi Yang, and Paridhi Gulati for their assistances with HPLC-DAD and proximate composition analyses, and Dr. Sathaporn Onanong for his assistance with LC-MS/MS. References Arráez-Román, D., Fu, S., Sawalha, S., Segura-Carretero, A., & Fernández-Gutiérrez, A. (2010). HPLC/CE-ESI-TOF-MS methods for the characterization of polyphenols in almond-skin extracts. Electrophoresis, 31(13), 2289–2296. Arslan, Y., Takaç, S., & Eken-Saraçoğlu, N. (2012). Kinetic study of hemicellulosic sugar production from hazelnut shells. Chemical Engineering Journal, 185, 23–28. Association of Official Analytical Chemists (AOAC) (2000). Official methods of analysis of AOAC international (17th ed.). Gaithersburg, MD, USA: AOAC International Retrieved from < http://www.nhbs.com/title/102701/official-methods-of-analysis-of-aoacinternational-17th-edition > . Benzie, I. F., & Strain, J. (1996). The ferric reducing ability of plasma (FRAP) as a measure of “antioxidant power”: The FRAP assay. Analytical Biochemistry, 239(1), 70–76. Beyer, K., Grishina, G., Bardina, L., Grishin, A., & Sampson, H. A. (2002). Identification of an 11S globulin as a major hazelnut food allergen in hazelnut-induced systemic reactions. Journal of Allergy and Clinical Immunology, 110(3), 517–523. Caglar, A., & Aydinli, B. (2009). Isothermal co-pyrolysis of hazelnut shell and ultra-high molecular weight polyethylene: The effect of temperature and composition on the amount of pyrolysis products. Journal of Analytical and Applied Pyrolysis, 86(2), 304–309. Ciarmiello, L. F., Mazzeo, M. F., Minasi, P., Peluso, A., De Luca, A., Piccirillo, P., et al. (2014). Analysis of different european hazelnut (Corylus avellana L.) cultivars: Authentication, phenotypic features, and phenolic profiles. Journal of Agricultural and Food Chemistry, 62(26), 6236–6246. Ciemniewska-Żytkiewicz, H., Verardo, V., Pasini, F., Bryś, J., Koczoń, P., & Caboni, M. F. (2015). Determination of lipid and phenolic fraction in two hazelnut (Corylus avellana L.) cultivars grown in Poland. Food Chemistry, 168, 615–622. Contini, M., Baccelloni, S., Massantini, R., & Anelli, G. (2008). Extraction of natural antioxidants from hazelnut (Corylus avellana L.) shell and skin wastes by long maceration at room temperature. Food Chemistry, 110(3), 659–669. Del Rio, D., Calani, L., Dall’Asta, M., & Brighenti, F. (2011). Polyphenolic composition of hazelnut skin. Journal of Agricultural and Food Chemistry, 59(18), 9935–9941. Demirbaş, A. (1999). Properties of charcoal derived from hazelnut shell and the

15